Mobile phone manufacturers’ accumulation in sensors, imaging algorithms and system scheduling can be transferred to the field of robotics. On April 19, at the 2026 Beijing Yizhuang Half Marathon and Humanoid Robot Half Marathon, humanoid robots began to collectively approach the finishing pace of human runners.
In the competition that day, the Honor humanoid robot "Lightning" of the Monkey King team won the championship in the robot category with a time of 50 minutes and 26 seconds (net time). On-site timing showed that the first human half-marathon runner crossed the finish line between about 1 hour and 07 minutes and 1 hour and 08 minutes. This means that the humanoid robot champion in this event has surpassed the finishing time of the human champion, with a difference of about 17 minutes.
Compared with last year's robot champion's time of about 2 hours and 40 minutes, the time of this year's top-ranked robot competition has been greatly reduced.
"This robot event is not a competition, but an industry roadshow." Liu Xingliang, director of the DCCI Internet Research Institute, told China Business News that this robot half-marathon also marks that 2026 will become an important node for humanoid robots to shift from scientific research to commercial deployment.
Robot runs half marathon in under 1 hour
At 7:30 on April 19, the 2026 Beijing Yizhuang Humanoid Robot Half Marathon started. More than 100 teams and about 300 robots from home and abroad participated in the competition, completing a 21.0975-kilometer race on a track that included ramps, complex roads, and nearly 90-degree turns.
In the end, Honor "Lightning" won the championship with a time of 50 minutes and 26 seconds, followed by the Thunder Lightning Team and Catching Fire Team with 50 minutes, 56 seconds and 53 minutes and 01 seconds respectively. The top three all used autonomous navigation to complete the game, and the overall results have entered the range of less than one hour.
Compared with the first event, the rules and difficulty of this competition have been improved. According to the reporter’s understanding, this year’s event introduces dual modes of “autonomous navigation” and “remote control”, and unifies the ranking through weighting; the track integrates more than 10 types of terrain, which puts forward higher requirements for path planning, dynamic balance and endurance. About 40% of the teams chose completely autonomous navigation, a significant increase compared with last year.
"This track covers a variety of complex road conditions and places extremely high requirements on the robot's motion stability, endurance and heat dissipation performance. It is a comprehensive test of the technical strength of humanoid robots," an Honor insider told reporters.
Judging from the participating teams, Beichuang's "Tiangong Ultra", Yushu "H1" and other models competed on the same field. Many teams completed the entire course, and the overall completion rate and stability were improved compared with the first session. During the competition, perception, decision-making, control and battery life were put to a continuous test in a real environment.
Liu Xingliang told reporters that the robot industry is undergoing three changes: First, it is moving from remote control to autonomy, and robots are gradually equipped with intelligent body capabilities. Second, it is moving from laboratory to engineering, from "can it run" to "whether it can operate stably". Third, it is moving from single-point technology to system capabilities, and perception, decision-making and control are beginning to be integrated and verified.
According to Li Qingdu, founder of Shanghai Zhuoyide Robot Co., Ltd., whether it is a running robot or a landing-oriented product, there is a consensus on joint design and appearance. He believes that judging from the marathon event, the differences in robotics technology at the current stage are more reflected in engineering optimization and system integration capabilities.
"After the humanoid robot hardware is ready, the greater opportunity lies in scene applications based on the hardware. This is the biggest opportunity in the future." Li Qingdu told reporters.
Terminal manufacturers are pouring into the robot track
As the competition results are quickly refreshed, participating robots have begun to demonstrate system-level engineering capabilities.
Taking the "Lightning" that won the championship this time as an example, the relevant person in charge of Honor told reporters that the robot is equipped with Honor's self-developed liquid cooling system. The liquid cooling pipes can penetrate deep into the motor like capillaries to take away heat. The high-power liquid pump can achieve a heat exchange flow rate of more than 4 liters per minute, which can solve the heat dissipation problem under high-load exercise conditions.
"The power system uses an integrated joint module with a peak torque of 400 Nm. At the control level, it relies on multi-sensor fusion and dynamic motion control algorithms to achieve real-time adaptation to complex road conditions and center of gravity control." the person in charge told reporters.
According to the reporter’s understanding, Honor’s current layout in the field of robotics has lasted for nearly two years. In the past two years, Honor has successively released robot-related positions, covering dexterous hand design, system software, embodied intelligent algorithms and test engineering, etc., and is promoted by the new industry incubation department. The company's management has stated that the robot business relies on its AI capabilities and terminal ecology to advance and is part of the overall strategy.
It is worth noting that on the hardware side, as more mobile phone manufacturers flood into the robotics track, the consumer electronics supply chain has also begun to extend to the industrial chain.
Lens Technology stated on its official platform on April 19 that in the Honor robot, Lens provides more than 130 core metal structural parts, covering key motion units such as heads, arms, hips and legs. Public information shows that the company’s robot-related business revenue has exceeded 1 billion yuan, of which core components have entered the supply chain of domestic and foreign leading customers. It aims to double the scale of core components and complete machines in 2026.
Lingyi Intelligent Manufacturing also released relevant information after the Honor Robot won the championship, saying that the company provided a full set of structural parts and surface treatment technology for the above-mentioned robots and had delivered products in batches.
At the same time, terminal manufacturers’ investment in the field of robotics continues to increase. Taking Honor as an example, it has been continuously releasing robot-related positions since 2025, covering algorithms, structural design, system software and testing, and is promoted by the new industry incubation department. Manufacturers including Huawei, vivo, Xiaomi, and ZTE have also entered the robot track through self-research or cooperation.
The industry generally believes that mobile phone manufacturers’ accumulation in sensors, imaging algorithms and system scheduling can be transferred to the field of robots, and mature 3C supply chain capabilities will also help promote the scale and cost reduction of robot hardware.
However, robots are still a more complex industrial path. Capabilities such as mechanical structure, motion control, and electronic control systems are significantly different from traditional consumer electronics, and most manufacturers need to cooperate to make up for their shortcomings. At the same time, production line transformation, interdisciplinary talent integration, and longer R&D cycles also pose practical challenges.
On the commercialization path, Li Qingdu told reporters that as hardware gradually matures, the biggest opportunities will shift to scenario applications. “The key is to achieve the rapid implementation of high-value scenarios through AI on a stable and reliable platform.” He believes that more commercialization opportunities will emerge this year, but the core lies in verifying return on investment and real demand.
Liu Xingliang believes that commercialization needs to meet three points at the same time: stability, cost and scale capabilities, that is, whether it can operate stably in the long term, whether it is economical, and whether it can achieve batch delivery. From the perspective of application direction, in the short term, it will focus on industry and logistics, in the medium term, it will expand to commercial services, and in the long term, the space will still be in the home scene.